Re: [lmi] Headless GUI tests on corporate server

[Greg Chicares]

On 11/13/21 10:18 PM, Vadim Zeitlin wrote:
> On Sat, 13 Nov 2021 22:07:37 +0000 Greg Chicares <gchicares@sbcglobal.net> 
> wrote:
> 
> GC> On 11/13/21 5:37 PM, Greg Chicares wrote:
> GC> [...]
> GC> > I'll start a separate thread for server issues.
> GC> 
> GC> Here's a transcript of a session on our corporate server:
> [...]
> GC> Instead, we want the DISPLAY value specified in this command:
> GC>   Xvfb :1 -screen 0 1280x1024x8 &
> GC> and that seems to get us farther:
> GC> 
> GC> /opt/lmi/src/lmi[0]$DISPLAY=:1 wine /opt/lmi/bin/wx_test --ash_nazg 
> --data_path=/opt/lmi/data --test log_error
> GC> Assertion '"Assert failure"' failed
> GC> (incorrect bitmap type in wxBitmap::GetRawData()).
> GC> [file /opt/lmi/src/lmi/third_party/wx/src/msw/bitmap.cpp, line 1423]
> GC> ^C%
> GC> 
> GC> The string "incorrect bitmap type" doesn't occur anywhere in
> GC> this mailing list's two decades of archives. Vadim, have you
> GC> any idea what could cause that?
> 
>  Yes, it's certainly due to the "x8" specified above, as raw bitmap access
> is not supported in palette-using video modes. The main reason for this is
> that I thought nobody used them any longer, when I was implementing support
> for this (and that was 10+ years ago, so I think this is even truer today),
> but it also probably would be more difficult to support 8bpp than the other
> modes.
> 
>  So I'd advise to retry with the usual x24

TL;DR: Yes, indeed.

> instead, e.g. 1600x1200x24 that
> I use in the CI workflow file. Or is there some reason to use 8bpp that I'm
> forgetting about?

On my own machine, I had been using
  Xvfb :1 -screen 0 1280x960x24
But who else today uses a 4x3 aspect ratio, which is perfect
for doing actual work but not ideal for playing games and
watching movies? And who else uses 192 DPI, which is unpopular
though ideal? And why would this corporate server support 24-bit
color? Isn't 8-bit color the safest choice, because everything
since the 1980s would assuredly support it? That's what I was
thinking, so I decided to use "the default screen configuration"
from the manpage:

https://linux.die.net/man/1/xvfb
|
| Xvfb -pixdepths 3 27 -fbdir /var/tmp
|    The server will listen for connections as server number 0,
|    will have the default screen configuration (one screen,
|    1280x1024x8), will also support pixmap depths of 3 and 27,
|    and will use memory mapped files in /var/tmp for the
|    framebuffer. 

because, obviously, defaults must work robustly.

But if I instead follow your advice, everything just works,
as demonstrated in the wonderfully uninteresting session
transcript below (where, yes, I do want to skip those four
tests). Thanks so much. A major milestone has been reached.

/home/REDACTED[0]$schroot --chroot=chroot:lmi_bookworm_4
/home/REDACTED[0]$Xvfb :1 -screen 0 1280x1024x24 &
[1] 125857
/home/REDACTED[0]$x_process=125857
/home/REDACTED[0]$DISPLAY=:1 wine /opt/lmi/bin/wx_test --ash_nazg 
--data_path=/opt/lmi/data --test log_error
NOTE: starting the test suite
log_error: started
time=0ms (for log_error)
log_error: ok
time=3ms (for all tests)
SUCCESS: 1 test successfully completed.
/home/REDACTED[0]$DISPLAY=:1 wine /opt/lmi/bin/wx_test --ash_nazg 
--data_path=/opt/lmi/data
NOTE: starting the test suite
about_dialog_version: started
About dialog version string is '20211025T2123Z'.
time=2116ms (for about_dialog_version)
about_dialog_version: ok
benchmark_census: started
time=0ms (for benchmark_census)
benchmark_census: ok
calculation_summary: started
time=13073ms (for calculation_summary)
calculation_summary: ok
configurable_settings: started
configurable_settings: skipped (not running distribution tests)
create_open_census: started
time=280ms (for create_open_census)
create_open_census: ok
create_open_database: started
time=782ms (for create_open_database)
create_open_database: ok
create_open_gpt: started
time=846ms (for create_open_gpt)
create_open_gpt: ok
create_open_illustration: started
time=2850ms (for create_open_illustration)
create_open_illustration: ok
create_open_mec: started
time=754ms (for create_open_mec)
create_open_mec: ok
create_open_policy: started
time=494ms (for create_open_policy)
create_open_policy: ok
create_open_rounding: started
time=1369ms (for create_open_rounding)
create_open_rounding: ok
create_open_strata: started
time=371ms (for create_open_strata)
create_open_strata: ok
create_open_text: started
time=127ms (for create_open_text)
create_open_text: ok
default_input: started
default_input: skipped (not running distribution tests)
default_update: started
default_update: skipped (not running distribution tests)
expiry_dates: started
Expiry dates: begin=2450449 (1996-12-31), end=2472011 (2056-01-13)
time=1ms (for expiry_dates)
expiry_dates: ok
input_sequences: started
time=49098ms (for input_sequences)
input_sequences: ok
input_validation: started
input_validation: skipped (not running distribution tests)
log_error: started
time=0ms (for log_error)
log_error: ok
paste_census: started
time=8106ms (for paste_census)
paste_census: ok
pdf_census: started
time=2827ms (for pdf_census)
pdf_census: ok
pdf_illustration: started
time=4180ms (for pdf_illustration)
pdf_illustration: ok
validate_output_census: started
time=15468ms (for validate_output_census)
validate_output_census: ok
validate_output_illustration: started
time=3961ms (for validate_output_illustration)
validate_output_illustration: ok
validate_output_mec: started
time=332ms (for validate_output_mec)
validate_output_mec: ok
time=107113ms (for all tests)
SUCCESS: 21 tests successfully completed.
NOTE: 4 tests were skipped
/home/REDACTED[0]$