building fpm with ifort

[urbanjost]

Is there any interest in a version of fpm that can be built with the Intel ifort compiler? There is what I think is a bug in ifort that can be worked around, as demonstrated by the single-file build in
https://github.com/urbanjost/fpm-tools/blob/main/tools/bootstrap/fpm_intel.F90
but it we just want to stick with gfortran for building fpm itself I don't want to make the kludge (essentially, changing the USE FPM_TOML to a USE TOMLF for TOML_TABLE). If you want to see the error, look for the lines with !x! and switch which module TOML_TABLE is used from).

Other than it being useful during development of fpm to use multiple compilers and their associated tools (not a trivial advantage by itself) does anyone have a use-case like an environment where everything must be built from source and a recent gfortran(1) is not installed? That is, is anyone being held back from using fpm(1) that would be able to use it if it built with ifort(1)?

[arjenmarkus]

On Windows, gfortran is not always easy to install - not impossible, but there is a wealth of possibilities, whereas people may be more used to environments like MS Visual Studio that hide the details of the build process. Op do 15 jul. 2021 om 15:37 schreef urbanjost ***@***.***>:

…

Is there any interest in a version of fpm that can be built with the Intel ifort compiler? There is what I think is a bug in ifort that can be worked around, as demonstrated by the single-file build in https://github.com/urbanjost/fpm-tools/blob/main/tools/bootstrap/fpm_intel.F90 but it we just want to stick with gfortran for building fpm itself I don't want to make the kludge (essentially, changing the USE FPM_TOML to a USE TOMLF for TOML_TABLE). If you want to see the error, look for the lines with !x! and switch which module TOML_TABLE is used from). Other than it being useful during development of fpm to use multiple compilers and their associated tools (not a trivial advantage by itself) does anyone have a use-case like an environment where everything must be built from source and a recent gfortran(1) is not installed? That is, is anyone being held back from using fpm(1) that would be able to use it if it built with ifort(1)? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#520>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN6YR6HWQTQIDFTBTSWN2LTX3QCZANCNFSM5ANSJ3FA> .

[arjenmarkus]

I compiled the current version with Intel oneAPI and that worked fine. I have not tested it to any useful extent, but at least the first step was done. Op do 15 jul. 2021 om 15:53 schreef Arjen Markus ***@***.***

…

: On Windows, gfortran is not always easy to install - not impossible, but there is a wealth of possibilities, whereas people may be more used to environments like MS Visual Studio that hide the details of the build process. Op do 15 jul. 2021 om 15:37 schreef urbanjost ***@***.***>: > Is there any interest in a version of fpm that can be built with the > Intel ifort compiler? There is what I think is a bug in ifort that can be > worked around, as demonstrated by the single-file build in > > https://github.com/urbanjost/fpm-tools/blob/main/tools/bootstrap/fpm_intel.F90 > but it we just want to stick with gfortran for building fpm itself I > don't want to make the kludge (essentially, changing the USE FPM_TOML to a > USE TOMLF for TOML_TABLE). If you want to see the error, look for the lines > with !x! and switch which module TOML_TABLE is used from). > > Other than it being useful during development of fpm to use multiple > compilers and their associated tools (not a trivial advantage by itself) > does anyone have a use-case like an environment where everything must be > built from source and a recent gfortran(1) is not installed? That is, is > anyone being held back from using fpm(1) that would be able to use it if it > built with ifort(1)? > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#520>, or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAN6YR6HWQTQIDFTBTSWN2LTX3QCZANCNFSM5ANSJ3FA> > . >