# FPM - failing at the first hurdle **URL:** \ **Category:** Uncategorized\ **Created:** [June 15, 2026, 10:18am UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968 "2026-06-15T10:18:56Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![HighPerformanceMark](https://avatars.discourse-cdn.com/v4/letter/h/c37758/32.png) [@HighPerformanceMark](https://fortran-lang.discourse.group/u/HighPerformanceMark)\ **Post date:** [June 15, 2026, 10:18am UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/1 "2026-06-15T10:18:56Z")
I finally got round to moving away from Make and towards FPM. So I ran `port install fpm` (Yes, I use MacPorts) carried on following the instructions and executed `fpm new first_steps` then `fpm run` to which my Mac replied > first\_steps.f90 compiling… > [0%] Compiling…:failed command gfortran -c ././src/first\_steps.f90 -Wall -Wextra -fPIC -fmax-errors=1 -g -fcheck=bounds -fcheck=array-temps -fbacktrace -fcoarray=single -fimplicit-none -Werror=implicit-interface -ffree-form -J build/gfortran\_87E2AE0597D39913 -Ibuild/gfortran\_87E2AE0597D39913 -o build/gfortran\_87E2AE0597D39913/first\_steps/src\_first\_steps.f90.o\>build/gfortran\_87E2AE0597D39913/first\_steps/src\_first\_steps.f90.o.log 2\>&1 > _run_:Invalid command line > STOP 1 Which prompted the question - where did all those options for gfortran come from? I can’t find them anywhere in the fpm configuration files. Oh, yeah, the other question, what went wrong, well I can probably sort that out myself when I figure out a bit more about how fpm is going about its business. And yes, I have pored over the documentation. Thanks for any help offered. ---
**Author:** ![davidpfister](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/davidpfister/32/3949_2.png) [@davidpfister](https://fortran-lang.discourse.group/u/davidpfister)\ **Post date:** [June 15, 2026, 11:48am UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/2 "2026-06-15T11:48:26Z")
I guess you have to look into the source code to understand where the flags are coming from. Fpm is coming with two different hard-coded profiles: ‘release’ and ‘debug’. The flags are found in the [fpm\_compiler.f90](https://github.com/fortran-lang/fpm/blob/eaffbb36086abdb16c0d052961a3e7240cb22b0a/src/fpm_compiler.F90#L328) file. If you want to use you own profile and overwrite the default configuration, you can define a profile in the toml file, see the [doc](https://fpm.fortran-lang.org/spec/features.html). ```auto name = "my-package" version = "1.0.0" [features] # Release feature with optimization optim.flags = "-O3" optim.gfortran.flags = "-funroll-loops" optim.ifort.flags = "-unroll" optim.preprocess.cpp.macros = ["NDEBUG"] [profiles] optim = ["optim"] ``` Like this, you can create the profile you want and use it as: fpm run --profile optim ---
**Author:** ![FedericoPerini](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/federicoperini/32/1750_2.png) [@FedericoPerini](https://fortran-lang.discourse.group/u/FedericoPerini)\ **Post date:** [June 15, 2026, 1:54pm UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/3 "2026-06-15T13:54:45Z")
The message is genuinely misleading and it should be fixed imho. what’s `gfortran` named like on MacPorts? it would look like something like this should work: ```sh fpm run --compiler gfortran-mp-15 ``` or with the env variable: ```sh export FPM_FC=gfortran-mp-15 ``` ---
**Author:** ![HighPerformanceMark](https://avatars.discourse-cdn.com/v4/letter/h/c37758/32.png) [@HighPerformanceMark](https://fortran-lang.discourse.group/u/HighPerformanceMark)\ **Post date:** [June 16, 2026, 6:11am UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/4 "2026-06-16T06:11:03Z")
> [@FedericoPerini](#): > > `fpm run --compiler gfortran-mp-15` Thanks Federico That solved the problem. I do have `gfortran` aliased to point to `gfortran-mp-15` but I guess that fpm is using it’s own ways to find the compiler. Mark ---
**Author:** ![HighPerformanceMark](https://avatars.discourse-cdn.com/v4/letter/h/c37758/32.png) [@HighPerformanceMark](https://fortran-lang.discourse.group/u/HighPerformanceMark)\ **Post date:** [June 16, 2026, 6:15am UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/5 "2026-06-16T06:15:28Z")
Well, I did think, as I tried to figure this out “Maybe the compiler options are hard-coded into the fpm sources?” and then I thought … … that would be a really odd design decision, surely not? Thanks for the input Mark ---
**Author:** ![Machalot](https://avatars.discourse-cdn.com/v4/letter/m/ebca7d/32.png) [@Machalot](https://fortran-lang.discourse.group/u/Machalot)\ **Post date:** [June 16, 2026, 4:39pm UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/6 "2026-06-16T16:39:58Z")
> [@HighPerformanceMark](#): > > Well, I did think, as I tried to figure this out “Maybe the compiler options are hard-coded into the fpm sources?” and then I thought … > > … that would be a really odd design decision, surely not? It does make one wonder why the default options are not exposed in a configuration file. I’ve never developed software like this so I am not here to second guess, but I’m curious. ---
**Author:** ![FedericoPerini](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/federicoperini/32/1750_2.png) [@FedericoPerini](https://fortran-lang.discourse.group/u/FedericoPerini)\ **Post date:** [June 16, 2026, 7:31pm UTC](https://fortran-lang.discourse.group/t/fpm-failing-at-the-first-hurdle/10968/7 "2026-06-16T19:31:32Z")
Fully customizable and named sets of flags are now possible via features / profiles since v0.13.0 > **[Features and Profiles — Fortran Package Manager](https://fpm.fortran-lang.org/spec/features.html?highlight=profiles)** I’ll be the first to admit the original design has some friction re compiler/os/customization support: the default flags were effectively baked in, which is exactly the issue you’re running into. That’s part of why I pushed the features/profiles work into 0.13.0: it’s a substantial improvement and should give you the configurability you’re after. For the historical context behind the original choices, you can look up the repository issues section, e.g.: > > > How should we organize/specify compiler flags? > > Should we only have two sets o…f compiler flags? \`devel-options\` and \`release-options\`, and they are used to build all dependencies and executables? > > Also, should we come up with our own keywords for specifying the compiler options, and have a look up table to determine the correct flag for each compiler?