Meson can't recover from deletion of generated QAPI file(s)

Meson can't recover from deletion of generated QAPI file(s)

[Markus Armbruster]

Watch this:

    $ rm qapi/qapi-types-error.h 
    $ make
    Generating qemu-version.h with a meson_exe.py custom command
    Compiling C object qom/libqom.fa.p/object_interfaces.c.o
    In file included from ../qom/object_interfaces.c:4:
    /work/armbru/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
      275 | #include "qapi/qapi-types-error.h"
          |          ^~~~~~~~~~~~~~~~~~~~~~~~~
    compilation terminated.
    make: *** [Makefile.ninja:348: qom/libqom.fa.p/object_interfaces.c.o] Error 1

To recover, I have to run qapi-gen.py manually and exactly right, or
blow away the build tree and start over.

The old build system did what a build system should: it remade the files
that are missing or out of date.

I'm still too clueless about Meson to debug this, but I hope I can learn
from watching you fix it.

Re: Meson can't recover from deletion of generated QAPI file(s)

[Philippe Mathieu-Daudé]

On 9/9/20 2:37 PM, Markus Armbruster wrote:
> Watch this:
> 
>     $ rm qapi/qapi-types-error.h 
>     $ make
>     Generating qemu-version.h with a meson_exe.py custom command
>     Compiling C object qom/libqom.fa.p/object_interfaces.c.o
>     In file included from ../qom/object_interfaces.c:4:
>     /work/armbru/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
>       275 | #include "qapi/qapi-types-error.h"
>           |          ^~~~~~~~~~~~~~~~~~~~~~~~~
>     compilation terminated.
>     make: *** [Makefile.ninja:348: qom/libqom.fa.p/object_interfaces.c.o] Error 1
> 
> To recover, I have to run qapi-gen.py manually and exactly right, or
> blow away the build tree and start over.
> 
> The old build system did what a build system should: it remade the files
> that are missing or out of date.
> 
> I'm still too clueless about Meson to debug this, but I hope I can learn
> from watching you fix it.

Maybe related:
https://www.mail-archive.com/qemu-devel@nongnu.org/msg736135.html
and later:
https://www.mail-archive.com/qemu-devel@nongnu.org/msg738777.html

Re: Meson can't recover from deletion of generated QAPI file(s)

[Claudio Fontana]

On 9/9/20 3:06 PM, Philippe Mathieu-Daudé wrote:
> On 9/9/20 2:37 PM, Markus Armbruster wrote:
>> Watch this:
>>
>>     $ rm qapi/qapi-types-error.h 
>>     $ make
>>     Generating qemu-version.h with a meson_exe.py custom command
>>     Compiling C object qom/libqom.fa.p/object_interfaces.c.o
>>     In file included from ../qom/object_interfaces.c:4:
>>     /work/armbru/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
>>       275 | #include "qapi/qapi-types-error.h"
>>           |          ^~~~~~~~~~~~~~~~~~~~~~~~~
>>     compilation terminated.
>>     make: *** [Makefile.ninja:348: qom/libqom.fa.p/object_interfaces.c.o] Error 1
>>
>> To recover, I have to run qapi-gen.py manually and exactly right, or
>> blow away the build tree and start over.
>>
>> The old build system did what a build system should: it remade the files
>> that are missing or out of date.
>>
>> I'm still too clueless about Meson to debug this, but I hope I can learn
>> from watching you fix it.
> 
> Maybe related:
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg736135.html
> and later:
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg738777.html
> 

My use case is this:
my simple build script uses git rebase -x to run a build for each commit of a series, and fails if at any point one of the commits breaks the build or tests.

FILE ~/bin/check-build.sh:

#! /bin/bash
git rebase $1 -x build.sh

FILE ~/bin/build.sh:

#! /bin/bash
set -x
set -e

rm -fr build-tcg
mkdir build-tcg
cd build-tcg
../configure --enable-tcg --disable-kvm --disable-hax
make -j120
make -j120 check
cd ..

rm -fr build-nontcg
mkdir build-nontcg
cd build-nontcg
../configure --disable-tcg --enable-kvm --enable-hax
make -j120
make -j120 check
cd ..

rm -fr build-all
mkdir build-all
cd build-all
../configure --enable-tcg --enable-kvm --enable-hax
make -j120
make -j120 check
cd ..

--------------------

In case it helps with reproducing the problem.

Ciao,

Claudio

Re: Meson can't recover from deletion of generated QAPI file(s)

[Marc-André Lureau]

[-- Attachment #1: Type: text/plain, Size: 1165 bytes --]

Hi

On Wed, Sep 9, 2020 at 4:37 PM Markus Armbruster <armbru@redhat.com> wrote:

> Watch this:
>
>     $ rm qapi/qapi-types-error.h
>     $ make
>     Generating qemu-version.h with a meson_exe.py custom command
>     Compiling C object qom/libqom.fa.p/object_interfaces.c.o
>     In file included from ../qom/object_interfaces.c:4:
>     /work/armbru/qemu/include/qapi/error.h:275:10: fatal error:
> qapi/qapi-types-error.h: No such file or directory
>       275 | #include "qapi/qapi-types-error.h"
>           |          ^~~~~~~~~~~~~~~~~~~~~~~~~
>     compilation terminated.
>     make: *** [Makefile.ninja:348: qom/libqom.fa.p/object_interfaces.c.o]
> Error 1
>
> To recover, I have to run qapi-gen.py manually and exactly right, or
> blow away the build tree and start over.
>
> The old build system did what a build system should: it remade the files
> that are missing or out of date.
>
> I'm still too clueless about Meson to debug this, but I hope I can learn
> from watching you fix it.
>

It looks like the ninja build handles that fine, so this is most likely a
ninja2make conversion issue.


-- 
Marc-André Lureau

[-- Attachment #2: Type: text/html, Size: 1633 bytes --]

Re: Meson can't recover from deletion of generated QAPI file(s)

[Paolo Bonzini]

On 09/09/20 14:37, Markus Armbruster wrote:
> Watch this:
> 
>     $ rm qapi/qapi-types-error.h 
>     $ make
>     Generating qemu-version.h with a meson_exe.py custom command
>     Compiling C object qom/libqom.fa.p/object_interfaces.c.o
>     In file included from ../qom/object_interfaces.c:4:
>     /work/armbru/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
>       275 | #include "qapi/qapi-types-error.h"
>           |          ^~~~~~~~~~~~~~~~~~~~~~~~~
>     compilation terminated.
>     make: *** [Makefile.ninja:348: qom/libqom.fa.p/object_interfaces.c.o] Error 1
> 
> To recover, I have to run qapi-gen.py manually and exactly right, or
> blow away the build tree and start over.
> 
> The old build system did what a build system should: it remade the files
> that are missing or out of date.
> 
> I'm still too clueless about Meson to debug this, but I hope I can learn
> from watching you fix it.

Unfortunately you won't learn much about Meson, you would learn that 
Make is messy but you don't need that.  You can also learn a little bit 
about the new design of the QEMU build system though, so I'll explain 
and not just send a patch.

The bad news this tells you about the build system is that, when 
debugging an issue, you have to figure out if it's a bug in Meson,
in the meson.build files, or in ninja2make.  Of course the second
is the common case, but you never know especially now that there are
more people using ninja2make in anger.

Generating Makefile.ninja gets finicky because ninja (while it has other
things I don't like) is a little more expressive than Make as far as
simple build rules are concerned, therefore it doesn't need the stamp
file trick.  So while we there may be one or two more bugs like this one
down the road, ninja2make should not be an issue as soon as its teething
problems are solved.

(As an aside: the GNU Make 4.3 "grouped targets" feature can sometimes
eliminate stamp files, but it would not help here.  The stamp file has
another feature, namely the custom command can decide not to touch
its outputs if they won't change.  This avoid more rebuilds.  Grouped
targets don't have a way to do with that).

The good news is that there's an easy(ish) way to do this.  The build 
system is quite "linear" in how it works, so the first step should be to 
look at build.ninja and see what the rules are like.  Here you'd see
something like:

  build long list of files: CUSTOM_COMMAND actual prerequisites
   COMMAND = ...
   description = Generating$ shared$ QAPI$ source$ files

Your twenty-plus-years-of-writing-Makefiles spidey sense will tingle, as 
you can figure out that this is not going to be trivial to convert to 
Makefiles.  If you open Makefile.ninja you see the familiar stamp file
trick:

  long list of files: CUSTOM_COMMAND@57579de3eef.stamp; @:
  CUSTOM_COMMAND@57579de3eef.stamp: actual prerequisites
	$(ninja-command-restat)

and that's where the bug is.  If you delete one of the output files, Make
only runs ":" and does not rebuild it.  One solution is to add:

  ifneq (long list of files, $(wildcard long list of files))
  .PHONY: CUSTOM_COMMAND@57579de3eef.stamp
  endif

This way, if any of the prerequites is missing (not just older than the 
stamp file), the rule for CUSTOM_COMMAND@57579de3eef.stamp will always 
be executed.

This is fairly simple to do:

diff --git a/scripts/ninjatool.py b/scripts/ninjatool.py
index 627a1cab45..6f0e35c727 100755
--- a/scripts/ninjatool.py
+++ b/scripts/ninjatool.py
@@ -908,6 +908,9 @@ class Ninja2Make(NinjaParserEventsWithVars):
             else:
                 stamp = '%s@%s.stamp' % (rule, sha1_text(targets)[0:11])
             self.print('%s: %s; @:' % (targets, stamp))
+            self.print('ifneq (%s, $(wildcard %s))' % (targets, targets))
+            self.print('.PHONY: %s' % (stamp, ))
+            self.print('endif')
             self.print('%s: %s | %s; ${ninja-command-restat}' % (stamp, inputs, orderonly))
             self.rule_targets[rule].append(stamp)
             self.stamp_targets[rule].append(stamp)

To avoid this whole class of issues we could just use ninja to build QEMU
(Make would launch it, so there would still be no user-facing changes).
ninja2make's main strength was that it supported incremental conversion,
but right now all of the binaries are built by Meson therefore it's not
really *necessary* anymore.  Dropping ninja2make removes a relatively
expensive part of the build as well as a nontrivial amount of code.

Another advantage would be that ninja tracks command lines and automatically
rebuilds things if the command line has changed.  This is quite hard and
expensive to do with Make so ninja2make does not even try, but it has bitten
Philippe.

Of course, the main disadvantage is that it adds another dependency.

I will send the above patch formally soonish, but I wouldn't mind if somebody
else helped merging it.

Paolo

Re: Meson can't recover from deletion of generated QAPI file(s)

[Paolo Bonzini]

On 09/09/20 15:35, Claudio Fontana wrote:
> On 9/9/20 3:06 PM, Philippe Mathieu-Daudé wrote:
>> Maybe related:
>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg736135.html
>> and later:
>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg738777.html
>>
> 
> My use case is this:

Hi Claudio,

unlike Markus's, your issue is a genuine meson.build bug.  It should be
fixed like so:

diff --git a/tests/meson.build b/tests/meson.build
index 998e4c48f9..721641afbb 100644
--- a/tests/meson.build
+++ b/tests/meson.build
@@ -223,7 +223,7 @@ foreach test_name, extra: tests
     src += test_ss.all_sources()
     deps += test_ss.all_dependencies()
   endif
-  exe = executable(test_name, src, dependencies: deps)
+  exe = executable(test_name, src, genh, dependencies: deps)

   test(test_name, exe,
        depends: test_deps.get(test_name, []),


Paolo

Re: Meson can't recover from deletion of generated QAPI file(s)

[Claudio Fontana]

On 9/9/20 4:35 PM, Paolo Bonzini wrote:
> On 09/09/20 15:35, Claudio Fontana wrote:
>> On 9/9/20 3:06 PM, Philippe Mathieu-Daudé wrote:
>>> Maybe related:
>>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg736135.html
>>> and later:
>>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg738777.html
>>>
>>
>> My use case is this:
> 
> Hi Claudio,
> 
> unlike Markus's, your issue is a genuine meson.build bug.  It should be
> fixed like so:
> 
> diff --git a/tests/meson.build b/tests/meson.build
> index 998e4c48f9..721641afbb 100644
> --- a/tests/meson.build
> +++ b/tests/meson.build
> @@ -223,7 +223,7 @@ foreach test_name, extra: tests
>      src += test_ss.all_sources()
>      deps += test_ss.all_dependencies()
>    endif
> -  exe = executable(test_name, src, dependencies: deps)
> +  exe = executable(test_name, src, genh, dependencies: deps)
> 
>    test(test_name, exe,
>         depends: test_deps.get(test_name, []),
> 
> 
> Paolo
> 

Will test right away, thanks!

C

Re: Meson can't recover from deletion of generated QAPI file(s)

[Claudio Fontana]

On 9/9/20 4:36 PM, Claudio Fontana wrote:
> On 9/9/20 4:35 PM, Paolo Bonzini wrote:
>> On 09/09/20 15:35, Claudio Fontana wrote:
>>> On 9/9/20 3:06 PM, Philippe Mathieu-Daudé wrote:
>>>> Maybe related:
>>>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg736135.html
>>>> and later:
>>>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg738777.html
>>>>
>>>
>>> My use case is this:
>>
>> Hi Claudio,
>>
>> unlike Markus's, your issue is a genuine meson.build bug.  It should be
>> fixed like so:
>>
>> diff --git a/tests/meson.build b/tests/meson.build
>> index 998e4c48f9..721641afbb 100644
>> --- a/tests/meson.build
>> +++ b/tests/meson.build
>> @@ -223,7 +223,7 @@ foreach test_name, extra: tests
>>      src += test_ss.all_sources()
>>      deps += test_ss.all_dependencies()
>>    endif
>> -  exe = executable(test_name, src, dependencies: deps)
>> +  exe = executable(test_name, src, genh, dependencies: deps)
>>
>>    test(test_name, exe,
>>         depends: test_deps.get(test_name, []),
>>
>>
>> Paolo
>>
> 
> Will test right away, thanks!
> 
> C
> 

Hmm, I still encounter problems:

Passed all 120 iotests
+ cd ..
+ rm -fr build-nontcg
+ mkdir build-nontcg
+ cd build-nontcg
+ ../configure --disable-tcg --enable-kvm --enable-hax
cross containers  no

NOTE: guest cross-compilers enabled: cc
The Meson build system
Version: 0.55.1
Source dir: /dev/shm/cfontana/qemu
Build dir: /dev/shm/cfontana/qemu/build-nontcg
Build type: native build
Project name: qemu
Project version: 5.1.50
C compiler for the host machine: cc (gcc 7.5.0 "cc (SUSE Linux) 7.5.0")
C linker for the host machine: cc ld.bfd 2.34.0.20200325-386
Host machine cpu family: x86_64
Host machine cpu: x86_64
../meson.build:10: WARNING: Module unstable-keyval has no backwards or forwards compatibility and might not exist in future releases.
Program sh found: YES
Program python3 found: YES (/usr/bin/python3)
Configuring ninjatool using configuration
C++ compiler for the host machine: c++ (gcc 7.5.0 "c++ (SUSE Linux) 7.5.0")
C++ linker for the host machine: c++ ld.bfd 2.34.0.20200325-386
Library m found: YES
Library util found: YES
Found pkg-config: /usr/bin/pkg-config (0.29.2)
Run-time dependency pixman-1 found: YES 0.34.0
Library pam found: YES
Library aio found: YES
Run-time dependency zlib found: YES 1.2.11
Run-time dependency xkbcommon found: NO (tried pkgconfig)
Library rt found: YES
sdl2-config found: NO
Run-time dependency sdl2 found: NO (tried pkgconfig and config-tool)
Run-time dependency libpng found: NO (tried pkgconfig)
Has header "jpeglib.h" : NO 
Has header "sasl/sasl.h" : YES 
Library sasl2 found: YES
Run-time dependency u2f-emu found: NO (tried pkgconfig)
Run-time dependency libkeyutils found: NO (tried pkgconfig)
Checking for function "gettid" : NO 
Configuring config-host.h using configuration
Program scripts/minikconf.py found: YES
Configuring i386-softmmu-config-target.h using configuration
Configuring i386-softmmu-config-devices.mak with command
Reading depfile: /dev/shm/cfontana/qemu/build-nontcg/meson-private/i386-softmmu-config-devices.mak.d
Configuring i386-softmmu-config-devices.h using configuration
Configuring x86_64-softmmu-config-target.h using configuration
Configuring x86_64-softmmu-config-devices.mak with command
Reading depfile: /dev/shm/cfontana/qemu/build-nontcg/meson-private/x86_64-softmmu-config-devices.mak.d
Configuring x86_64-softmmu-config-devices.h using configuration
Program scripts/hxtool found: YES
Program scripts/shaderinclude.pl found: YES
Program scripts/qapi-gen.py found: YES
Program scripts/qemu-version.sh found: YES
Run-time dependency threads found: YES
Program keycodemapdb/tools/keymap-gen found: YES
Program scripts/decodetree.py found: YES
Program ../scripts/modules/module_block.py found: YES
Program nm found: YES
Program scripts/undefsym.py found: YES
Program scripts/feature_to_c.sh found: YES
Program bzip2 found: YES
Configuring 50-edk2-i386-secure.json using configuration
Configuring 50-edk2-x86_64-secure.json using configuration
Configuring 60-edk2-aarch64.json using configuration
Configuring 60-edk2-arm.json using configuration
Configuring 60-edk2-i386.json using configuration
Configuring 60-edk2-x86_64.json using configuration
Program qemu-keymap found: YES
Program python3 found: YES (/usr/bin/python3)
Program diff found: YES
Program dbus-daemon found: YES
Program initrd-stress.sh found: YES
Build targets in project: 410

qemu 5.1.50

            Install prefix: /usr/local
            BIOS directory: /usr/local/share/qemu
             firmware path: /usr/local/share/qemu-firmware
          binary directory: /usr/local/bin
         library directory: /usr/local/lib
          module directory: /usr/local/lib/qemu
         libexec directory: /usr/local/libexec
         include directory: /usr/local/include
          config directory: /usr/local/etc
     local state directory: /usr/local/var
          Manual directory: share/man
             Doc directory: /usr/local/share/doc
           Build directory: /dev/shm/cfontana/qemu/build-nontcg
               Source path: /dev/shm/cfontana/qemu
                GIT binary: git
            GIT submodules: ui/keycodemapdb tests/fp/berkeley-testfloat-3 tests/fp/berkeley-softfloat-3 meson capstone slirp
                C compiler: cc
           Host C compiler: cc
              C++ compiler: c++
                   ARFLAGS: rv
                    CFLAGS: -O2 -g -fPIE -DPIE -std=gnu99 -Wall
               QEMU_CFLAGS: -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 -m64 -mcx16 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -Wstrict-prototypes -Wredundant-decls -Wundef -Wwrite-strings -Wmissing-prototypes -fno-strict-aliasing -fno-common -fwrapv  -Wold-style-declaration -Wold-style-definition -Wtype-limits -Wformat-security -Wformat-y2k -Winit-self -Wignored-qualifiers -Wempty-body -Wnested-externs -Wendif-labels -Wexpansion-to-defined -Wno-missing-include-dirs -Wno-shift-negative-value -Wno-psabi -fstack-protector-strong
              QEMU_LDFLAGS: -Wl,--warn-common -Wl,-z,relro -Wl,-z,now -m64  -fstack-protector-strong
                      make: make
                    python: /usr/bin/python3 (version: 3.6)
              sphinx-build: 
               genisoimage: /usr/bin/mkisofs
             slirp support: YES
                      smbd: "/usr/sbin/smbd"
            module support: NO
                  host CPU: x86_64
           host endianness: little
               target list: i386-softmmu x86_64-softmmu
             gprof enabled: NO
            sparse enabled: NO
            strip binaries: YES
                  profiler: NO
              static build: YES
               SDL support: NO
         SDL image support: NO
               GTK support: NO
            GTK GL support: NO
                    pixman: YES
               VTE support: NO
              TLS priority: "NORMAL"
            GNUTLS support: NO
                 libgcrypt: NO
                    nettle: NO
                  libtasn1: NO
                       PAM: YES
             iconv support: YES
            curses support: YES
             virgl support: NO
              curl support: NO
           mingw32 support: NO
             Audio drivers: oss
      Block whitelist (rw): 
      Block whitelist (ro): 
            VirtFS support: NO
         Multipath support: NO
               VNC support: YES
          VNC SASL support: YES
          VNC JPEG support: NO
           VNC PNG support: NO
               xen support: NO
            brlapi support: NO
             Documentation: NO
                       PIE: YES
               vde support: NO
            netmap support: NO
         Linux AIO support: YES
    Linux io_uring support: NO
        ATTR/XATTR support: YES
             Install blobs: YES
       malloc trim support: YES
              RDMA support: NO
            PVRDMA support: NO
               fdt support: NO
                membarrier: NO
            preadv support: YES
                 fdatasync: YES
                   madvise: YES
             posix_madvise: YES
            posix_memalign: YES
         libcap-ng support: NO
         vhost-net support: YES
      vhost-crypto support: YES
        vhost-scsi support: YES
       vhost-vsock support: YES
        vhost-user support: YES
     vhost-user-fs support: YES
        vhost-vdpa support: YES
            Trace backends: log
             spice support: NO
               rbd support: NO
            xfsctl support: NO
         smartcard support: NO
               U2F support: NO
                    libusb: NO
             usb net redir: NO
            OpenGL support: NO
            OpenGL dmabufs: NO
          libiscsi support: NO
            libnfs support: NO
         build guest agent: YES
           seccomp support: NO
         coroutine backend: ucontext
            coroutine pool: YES
         debug stack usage: NO
           mutex debugging: NO
              crypto afalg: NO
         GlusterFS support: NO
                      gcov: NO
               TPM support: YES
            libssh support: NO
             QOM debugging: YES
      Live block migration: YES
               lzo support: NO
            snappy support: NO
             bzip2 support: NO
             lzfse support: NO
              zstd support: NO
         NUMA host support: NO
                   libxml2: NO
          tcmalloc support: NO
          jemalloc support: NO
         avx2 optimization: YES
      avx512f optimization: NO
       replication support: YES
             bochs support: YES
             cloop support: YES
               dmg support: YES
           qcow v1 support: YES
               vdi support: YES
             vvfat support: YES
               qed support: YES
         parallels support: YES
          sheepdog support: YES
                  capstone: YES
           libpmem support: NO
         libdaxctl support: NO
                   libudev: NO
           default devices: YES
            plugin support: NO
           fuzzing support: NO
                       gdb: /usr/bin/gdb
          thread sanitizer: NO
                  rng-none: NO
             Linux keyring: YES
Found ninja-1.8.2 at /usr/bin/ninja
+ make -j120
/usr/bin/python3 -B /dev/shm/cfontana/qemu/meson/meson.py introspect --tests --benchmarks | /usr/bin/python3 -B scripts/mtest2make.py > Makefile.mtest
./ninjatool -t ninja2make --omit clean dist uninstall cscope TAGS ctags < build.ninja > Makefile.ninja
make[1]: Entering directory '/dev/shm/cfontana/qemu/slirp'

[...]

  CC      pvh_main.o
  BUILD   multiboot.img
  BUILD   linuxboot.img
  BUILD   linuxboot_dma.img
  BUILD   kvmvapic.img
  BUILD   multiboot.raw
  BUILD   linuxboot.raw
  BUILD   kvmvapic.raw
  BUILD   linuxboot_dma.raw
  SIGN    multiboot.bin
  SIGN    kvmvapic.bin
  SIGN    linuxboot.bin
  SIGN    linuxboot_dma.bin
Linking target contrib/ivshmem-client/ivshmem-client
  BUILD   pvh.img
  BUILD   pvh.raw
In file included from tests/qapi-builtin-visit.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
 #include "qapi/qapi-types-error.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/../test-qapi-events-sub-sub-module.h:17:0,
                 from tests/include/test-qapi-events-sub-module.h:16,
                 from tests/test-qapi-events.h:16,
                 from tests/test-qapi-events.c:15:
tests/include/../test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from /dev/shm/cfontana/qemu/include/qapi/dealloc-visitor.h:17:0,
                 from tests/qapi-builtin-types.c:14:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from /dev/shm/cfontana/qemu/include/qapi/dealloc-visitor.h:17:0,
                 from tests/test-qapi-types.c:14:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
  SIGN    pvh.bin
In file included from /dev/shm/cfontana/qemu/include/qapi/dealloc-visitor.h:17:0,
                 from tests/include/test-qapi-types-sub-module.c:14:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/test-qapi-visit-sub-module.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
 #include "qapi/qapi-types-error.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1760: tests/libtestqapi.a.p/meson-generated_.._qapi-builtin-types.c.o] Error 1
make: *** Waiting for unfinished jobs....
make: *** [Makefile.ninja:1761: tests/libtestqapi.a.p/meson-generated_.._qapi-builtin-visit.c.o] Error 1
make: *** [Makefile.ninja:1766: tests/libtestqapi.a.p/meson-generated_.._test-qapi-events.c.o] Error 1
make: *** [Makefile.ninja:1770: tests/libtestqapi.a.p/meson-generated_.._test-qapi-types.c.o] Error 1
make: *** [Makefile.ninja:1775: tests/libtestqapi.a.p/meson-generated_.._include_test-qapi-types-sub-module.c.o] Error 1
make: *** [Makefile.ninja:1776: tests/libtestqapi.a.p/meson-generated_.._include_test-qapi-visit-sub-module.c.o] Error 1

In file included from tests/test-qapi-visit-sub-sub-module.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
 #include "qapi/qapi-types-error.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/test-qapi-commands-sub-module.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from /dev/shm/cfontana/qemu/include/qapi/qmp/qlit.h:17:0,
                 from tests/test-qapi-introspect.h:15,
                 from tests/test-qapi-introspect.c:13:
/dev/shm/cfontana/qemu/include/qapi/qmp/qobject.h:35:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from /dev/shm/cfontana/qemu/include/qapi/dealloc-visitor.h:17:0,
                 from tests/test-qapi-types-sub-sub-module.c:14:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1768: tests/libtestqapi.a.p/meson-generated_.._test-qapi-introspect.c.o] Error 1
make: *** [Makefile.ninja:1773: tests/libtestqapi.a.p/meson-generated_.._include_test-qapi-commands-sub-module.c.o] Error 1
make: *** [Makefile.ninja:1771: tests/libtestqapi.a.p/meson-generated_.._test-qapi-visit-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-commands.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/../test-qapi-events-sub-sub-module.h:17:0,
                 from tests/include/test-qapi-events-sub-module.h:16,
                 from tests/include/test-qapi-events-sub-module.c:15:
tests/include/../test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1769: tests/libtestqapi.a.p/meson-generated_.._test-qapi-types-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-events-sub-sub-module.h:17:0,
                 from tests/test-qapi-events-sub-sub-module.c:15:
tests/test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1774: tests/libtestqapi.a.p/meson-generated_.._include_test-qapi-events-sub-module.c.o] Error 1
make: *** [Makefile.ninja:1763: tests/libtestqapi.a.p/meson-generated_.._test-qapi-commands.c.o] Error 1
make: *** [Makefile.ninja:1765: tests/libtestqapi.a.p/meson-generated_.._test-qapi-events-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-visit.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
 #include "qapi/qapi-types-error.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/test-qapi-commands-sub-sub-module.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/../test-qapi-events-sub-sub-module.h:17:0,
                 from tests/include/test-qapi-events-sub-module.h:16,
                 from tests/include/test-qapi-events-sub-module.c:15:
tests/include/../test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1769: tests/libtestqapi.a.p/meson-generated_.._test-qapi-types-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-events-sub-sub-module.h:17:0,
                 from tests/test-qapi-events-sub-sub-module.c:15:
tests/test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1774: tests/libtestqapi.a.p/meson-generated_.._include_test-qapi-events-sub-module.c.o] Error 1
make: *** [Makefile.ninja:1763: tests/libtestqapi.a.p/meson-generated_.._test-qapi-commands.c.o] Error 1
make: *** [Makefile.ninja:1765: tests/libtestqapi.a.p/meson-generated_.._test-qapi-events-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-visit.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
 #include "qapi/qapi-types-error.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/test-qapi-commands-sub-sub-module.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/../test-qapi-events-sub-sub-module.h:17:0,
                 from tests/include/test-qapi-events-sub-module.h:16,
                 from tests/include/test-qapi-events-sub-module.c:15:
tests/include/../test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1769: tests/libtestqapi.a.p/meson-generated_.._test-qapi-types-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-events-sub-sub-module.h:17:0,
                 from tests/test-qapi-events-sub-sub-module.c:15:
tests/test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1774: tests/libtestqapi.a.p/meson-generated_.._include_test-qapi-events-sub-module.c.o] Error 1
make: *** [Makefile.ninja:1763: tests/libtestqapi.a.p/meson-generated_.._test-qapi-commands.c.o] Error 1
make: *** [Makefile.ninja:1765: tests/libtestqapi.a.p/meson-generated_.._test-qapi-events-sub-sub-module.c.o] Error 1
In file included from tests/test-qapi-visit.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
 #include "qapi/qapi-types-error.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/test-qapi-commands-sub-sub-module.c:14:0:
/dev/shm/cfontana/qemu/include/qapi/visitor.h:18:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
In file included from tests/include/../test-qapi-commands-sub-sub-module.h:16:0,
                 from tests/include/test-qapi-commands-sub-module.h:16,
                 from tests/test-qapi-commands.h:16,
                 from tests/test-qapi-init-commands.c:14:
tests/include/../test-qapi-types-sub-sub-module.h:16:10: fatal error: qapi/qapi-builtin-types.h: No such file or directory
 #include "qapi/qapi-builtin-types.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile.ninja:1772: tests/libtestqapi.a.p/meson-generated_.._test-qapi-visit.c.o] Error 1
make: *** [Makefile.ninja:1762: tests/libtestqapi.a.p/meson-generated_.._test-qapi-commands-sub-sub-module.c.o] Error 1
make: *** [Makefile.ninja:1767: tests/libtestqapi.a.p/meson-generated_.._test-qapi-init-commands.c.o] Error 1
warning: execution failed: build.sh
You can fix the problem, and then run

  git rebase --continue

Re: Meson can't recover from deletion of generated QAPI file(s)

[Paolo Bonzini]

On 09/09/20 17:05, Claudio Fontana wrote:
> Hmm, I still encounter problems:

And another:

diff --git a/tests/meson.build b/tests/meson.build
index 998e4c48f9..95789f43b3 100644
--- a/tests/meson.build
+++ b/tests/meson.build
@@ -56,7 +56,7 @@ test_qapi_files = custom_target('Test QAPI files',
 # perhaps change qapi_gen to replace / with _, like Meson itself does?
 subdir('include')
 
-libtestqapi = static_library('testqapi', sources: [test_qapi_files, test_qapi_outputs_extra])
+libtestqapi = static_library('testqapi', sources: [test_qapi_files, genh, test_qapi_outputs_extra])
 testqapi = declare_dependency(link_with: libtestqapi)
 
 testblock = declare_dependency(dependencies: [block], sources: 'iothread.c')

Paolo

Re: Meson can't recover from deletion of generated QAPI file(s)

[Claudio Fontana]

On 9/9/20 5:10 PM, Paolo Bonzini wrote:
> On 09/09/20 17:05, Claudio Fontana wrote:
>> Hmm, I still encounter problems:
> 
> And another:
> 
> diff --git a/tests/meson.build b/tests/meson.build
> index 998e4c48f9..95789f43b3 100644
> --- a/tests/meson.build
> +++ b/tests/meson.build
> @@ -56,7 +56,7 @@ test_qapi_files = custom_target('Test QAPI files',
>  # perhaps change qapi_gen to replace / with _, like Meson itself does?
>  subdir('include')
>  
> -libtestqapi = static_library('testqapi', sources: [test_qapi_files, test_qapi_outputs_extra])
> +libtestqapi = static_library('testqapi', sources: [test_qapi_files, genh, test_qapi_outputs_extra])
>  testqapi = declare_dependency(link_with: libtestqapi)
>  
>  testblock = declare_dependency(dependencies: [block], sources: 'iothread.c')
> 
> Paolo
> 

adding this one as well fixes it for me, thanks!

Tested-by: Claudio Fontana <cfontana@suse.de>

Re: Meson can't recover from deletion of generated QAPI file(s)

[Markus Armbruster]

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 09/09/20 14:37, Markus Armbruster wrote:
>> Watch this:
>> 
>>     $ rm qapi/qapi-types-error.h 
>>     $ make
>>     Generating qemu-version.h with a meson_exe.py custom command
>>     Compiling C object qom/libqom.fa.p/object_interfaces.c.o
>>     In file included from ../qom/object_interfaces.c:4:
>>     /work/armbru/qemu/include/qapi/error.h:275:10: fatal error: qapi/qapi-types-error.h: No such file or directory
>>       275 | #include "qapi/qapi-types-error.h"
>>           |          ^~~~~~~~~~~~~~~~~~~~~~~~~
>>     compilation terminated.
>>     make: *** [Makefile.ninja:348: qom/libqom.fa.p/object_interfaces.c.o] Error 1
>> 
>> To recover, I have to run qapi-gen.py manually and exactly right, or
>> blow away the build tree and start over.
>> 
>> The old build system did what a build system should: it remade the files
>> that are missing or out of date.
>> 
>> I'm still too clueless about Meson to debug this, but I hope I can learn
>> from watching you fix it.
>
> Unfortunately you won't learn much about Meson, you would learn that 
> Make is messy but you don't need that.  You can also learn a little bit 
> about the new design of the QEMU build system though, so I'll explain 
> and not just send a patch.

Appreciated!

> The bad news this tells you about the build system is that, when 
> debugging an issue, you have to figure out if it's a bug in Meson,
> in the meson.build files, or in ninja2make.  Of course the second
> is the common case, but you never know especially now that there are
> more people using ninja2make in anger.
>
> Generating Makefile.ninja gets finicky because ninja (while it has other
> things I don't like) is a little more expressive than Make as far as
> simple build rules are concerned, therefore it doesn't need the stamp
> file trick.  So while we there may be one or two more bugs like this one
> down the road, ninja2make should not be an issue as soon as its teething
> problems are solved.

A Make replacement must address its issues to be credible.

Make's traditional inability to directly express "rule updates multiple
files" and "rule may or may not actually change its target(s)" has
always been an issue.  Countless Make users have had to learn the stamp
file work-around, and how to cope with its drawbacks.  I distinctly
remember grappling with it back in the 90s.

> (As an aside: the GNU Make 4.3 "grouped targets" feature can sometimes
> eliminate stamp files, but it would not help here.  The stamp file has
> another feature, namely the custom command can decide not to touch
> its outputs if they won't change.  This avoid more rebuilds.

This is *essential* when generating headers.  Without it, touching any
QAPI module or any part of the QAPI generator recompiles pretty much
everything.

>                                                               Grouped
> targets don't have a way to do with that).

Yes.  Also, TIL grouped targets :)

> The good news is that there's an easy(ish) way to do this.  The build 
> system is quite "linear" in how it works, so the first step should be to 
> look at build.ninja and see what the rules are like.  Here you'd see
> something like:
>
>   build long list of files: CUSTOM_COMMAND actual prerequisites
>    COMMAND = ...
>    description = Generating$ shared$ QAPI$ source$ files
>
> Your twenty-plus-years-of-writing-Makefiles spidey sense will tingle, as 
> you can figure out that this is not going to be trivial to convert to 
> Makefiles.  If you open Makefile.ninja you see the familiar stamp file
> trick:
>
>   long list of files: CUSTOM_COMMAND@57579de3eef.stamp; @:
>   CUSTOM_COMMAND@57579de3eef.stamp: actual prerequisites
> 	$(ninja-command-restat)
>
> and that's where the bug is.  If you delete one of the output files, Make
> only runs ":" and does not rebuild it.  One solution is to add:
>
>   ifneq (long list of files, $(wildcard long list of files))
>   .PHONY: CUSTOM_COMMAND@57579de3eef.stamp
>   endif
>
> This way, if any of the prerequites is missing (not just older than the 
> stamp file), the rule for CUSTOM_COMMAND@57579de3eef.stamp will always 
> be executed.

Neat trick!

Now I remember the old Makefiles actually didn't recover all by
themselves either, because they didn't use this trick.  Manual recovery
was "easy", though: remove the stamp file.

I ran into the "can't recover from deletion of generated QAPI file(s)"
issue when I updated a silly script of mine that helps me diff them.  To
force regeneration, the script removes them (relying on ccache to keep
compilation time in check).  Pre-Meson, it removed the stamp files.  I
didn't see any in meson.build, so I dropped their removal, nothing
worked, and I got quite confused.

> This is fairly simple to do:
>
> diff --git a/scripts/ninjatool.py b/scripts/ninjatool.py
> index 627a1cab45..6f0e35c727 100755
> --- a/scripts/ninjatool.py
> +++ b/scripts/ninjatool.py
> @@ -908,6 +908,9 @@ class Ninja2Make(NinjaParserEventsWithVars):
>              else:
>                  stamp = '%s@%s.stamp' % (rule, sha1_text(targets)[0:11])
>              self.print('%s: %s; @:' % (targets, stamp))
> +            self.print('ifneq (%s, $(wildcard %s))' % (targets, targets))
> +            self.print('.PHONY: %s' % (stamp, ))
> +            self.print('endif')
>              self.print('%s: %s | %s; ${ninja-command-restat}' % (stamp, inputs, orderonly))
>              self.rule_targets[rule].append(stamp)
>              self.stamp_targets[rule].append(stamp)
>
> To avoid this whole class of issues we could just use ninja to build QEMU
> (Make would launch it, so there would still be no user-facing changes).
> ninja2make's main strength was that it supported incremental conversion,
> but right now all of the binaries are built by Meson therefore it's not
> really *necessary* anymore.  Dropping ninja2make removes a relatively
> expensive part of the build as well as a nontrivial amount of code.

Makes sense to me.

> Another advantage would be that ninja tracks command lines and automatically
> rebuilds things if the command line has changed.  This is quite hard and
> expensive to do with Make so ninja2make does not even try, but it has bitten
> Philippe.
>
> Of course, the main disadvantage is that it adds another dependency.

Yes.  Drop in the bucket?

We already depend on ninja-the-language, we just feed it to ninja2make
instead of ninja.

> I will send the above patch formally soonish, but I wouldn't mind if somebody
> else helped merging it.

Thank you very much for taking the time to explain!