[hotfix] no tag failure #7
Conversation
| ##### Move to `ctp7_modules` work area | ||
| ```sh | ||
| cd /path/to/ctp7_modules | ||
| export PETA_STAGE=/path/to/local/installation/opt/gem-peta-stage |
There was a problem hiding this comment.
Is this targeting legacy only releases? In the templated version of the ctp7_modules we are getting rid of $PETA_STAGE and exporting it here doesn't seems to be a proper move...
There was a problem hiding this comment.
The hotfix is general, the documentation update is for the new framework.
Point me to an issue where removal of PETA_STAGE is discussed and I can rework this, however, PETA_STAGE is currently used by all templated compatible code building (it is just not required to be set by the developer except in certain cases)
| * `PackagePath` is the base directory of the package, defaults to `$(ProjectPath)` | ||
| * `INSTALL_PREFIX` : where to install the project, defaults to `/`, may be overridden for non-system installation | ||
| * `INSTALL_PATH` : base directory of the installed project, defaults to `/opt/$(Project)`, **should not be overridden** | ||
| * `PETA_PATH` : base directory where cross-compilation libraries will be installed on the host, defaults to `/opt/gem-peta-stage`, **should not be overridden** |
There was a problem hiding this comment.
Why shouldn't PETA_PATH be overridden? Isn't it supposed the base directory for all boards filesystem roots?
There was a problem hiding this comment.
Because both PETA_PATH and INSTALL_PATH are intentionally designed to only provide a fixed directory structure within the package, while INSTALL_PREFIX is the thing that gives freedom of where to put this predefined package structure.
If you have a suggestion on wording to make this more clear, please do
There was a problem hiding this comment.
Ok, it makes senses. I currently use PETA_PATH as PETA_STAGE because... it works.
|
|
||
| * `PETA_STAGE` is the location where the compiler will look for headers and libraries when cross-compiling, defaults to `$(PETA_PATH)`, may be overridden if developing features in multiple projects simultaneously. |
There was a problem hiding this comment.
How should PETA_STAGE be overridden when working with multiple backends?
There was a problem hiding this comment.
In an appropriate future update to mfZynq.mk, the flags and INSTALL_PATH would be wrapped inside target specific conditionals
diff --git a/mfZynq.mk b/mfZynq.mk
index 286d030..ddefba7 100644
--- a/mfZynq.mk
+++ b/mfZynq.mk
@@ -1,13 +1,13 @@
-PETA_STAGE?=$(PETA_PATH)/$(TARGET_BOARD)
+PETA_STAGE?=$(PETA_PATH)
CFLAGS= -fomit-frame-pointer -pipe -fno-common -fno-builtin \
-Wall -std=c++14 \
-march=armv7-a -mfpu=neon -mfloat-abi=hard \
-mthumb-interwork -mtune=cortex-a9 \
-DEMBED -Dlinux -D__linux__ -Dunix -fPIC \
- --sysroot=$(PETA_STAGE)
+ --sysroot=$(PETA_STAGE)/$(TARGET_BOARD)
-LDFLAGS+=--sysroot=$(PETA_STAGE)
+LDFLAGS+=--sysroot=$(PETA_STAGE)/$(TARGET_BOARD)
INSTALL_PATH=/mnt/persistent/$(ShortProject)Thus, the only developer change would be to point PETA_STAGE to the location where the compile time dependencies would live.
There was a problem hiding this comment.
Ok, maybe in a future update. Currently, I have to set PETA_PATH (i.e., PETA_STAGE?=$(PETA_PATH)/$(TARGET_BOARD)) to get the desired behavior.
As reported in cms-gem-daq-project/reedmuller-c#3, the script was failing to correctly parse the git info if a tag was not present, when using an older version of
gitThis was due to the fact that the shebang was set to fail on error, although the underlying issue had been fixed in the past.
This PR changes the shebang to remove the fail on error such that if no tag is present, a version of
0.0.0is assumed, and a release of0.0.X, where X is the number of commits since the initial commit, is made.Additionally, updates to the
READMEhave been made to better describe some of the scripts and variables, as well as adding an example section for the local installation use case.