You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have a concern about the header files that are exported publicly because they depends on some variables like:
HAVE_CONFIG_H
HAVE_STDINT_H
but these variables should be private to the library and the file config.h file should be private as well and not meant to be exported.
See for example the file "numbers.h".
In addition the usage of config.h and of the defines is essentially useless because the declarations in the header files are not actually adapted to the values each variable has.
In my understanding all the reference to these private variables should disappear from the exported header files and the config.h should not be included.
The text was updated successfully, but these errors were encountered:
franko
added a commit
to franko/libtexpdf
that referenced
this issue
Jan 3, 2023
Normally the test program should not use the private config variables of
the library but we are temporarily adding them to make the tests succed.
The problem was reported in the github issue:
sile-typesetter#27
Is this an artifact of the code originally being extracted from a larger upstream project where those might have been defined/used or just something that was never setup properly in the first place? One of the things that needs to be considered when fixing things like this is how we're going to keep this code in sync or at least refreshed occasionally with the upstream it was extracted/forked from.
I have a concern about the header files that are exported publicly because they depends on some variables like:
but these variables should be private to the library and the file config.h file should be private as well and not meant to be exported.
See for example the file "numbers.h".
In addition the usage of config.h and of the defines is essentially useless because the declarations in the header files are not actually adapted to the values each variable has.
In my understanding all the reference to these private variables should disappear from the exported header files and the config.h should not be included.
The text was updated successfully, but these errors were encountered: