what shell programming constructs can we count on in .bbclass files?

what shell programming constructs can we count on in .bbclass files?

[Robert P. J. Day]

  what shell (or level of POSIX compatibility) does one assume for
writing shell functions in .bbclass files?

  i'm looking at kernel-fitimage.bbclass and i see this:

  #
  # Step 7: Sign the image and add public key to U-Boot dtb
  #
  if [ "x${UBOOT_SIGN_ENABLE}" = "x1" ] ; then

surely we don't still need to use that ugly "x" prefix hack for string
comparisons, do we?

  a bit further down:

  do_assemble_fitimage_initramfs() {
        if echo ${KERNEL_IMAGETYPES} | grep -wq "fitImage" && \
                test -n "${INITRAMFS_IMAGE}" ; then

would it not be equivalent to write that second test as:

  [ -n "${INITRAMFS_IMAGE}"]

and so on. what are we allowed to count on?

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

Re: what shell programming constructs can we count on in .bbclass files?

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 1242 bytes --]

On 15 March 2017 at 10:44, Robert P. J. Day <rpjday@crashcourse.ca> wrote:

>   what shell (or level of POSIX compatibility) does one assume for
> writing shell functions in .bbclass files?
>
>   i'm looking at kernel-fitimage.bbclass and i see this:
>
>   #
>   # Step 7: Sign the image and add public key to U-Boot dtb
>   #
>   if [ "x${UBOOT_SIGN_ENABLE}" = "x1" ] ; then
>
> surely we don't still need to use that ugly "x" prefix hack for string
> comparisons, do we?
>
>   a bit further down:
>
>   do_assemble_fitimage_initramfs() {
>         if echo ${KERNEL_IMAGETYPES} | grep -wq "fitImage" && \
>                 test -n "${INITRAMFS_IMAGE}" ; then
>
> would it not be equivalent to write that second test as:
>
>   [ -n "${INITRAMFS_IMAGE}"]
>
> and so on. what are we allowed to count on?
>

xFOO = xBAR appears to be a long-standing hangover from the 70s, any POSIX
shell should be safe with quoted instead.

Our shell parser *should* be POSIX compliant.  There are a few places where
it isn't, but I believe those should be considered bugs.  At the end of the
day the shell is actually executed by /bin/sh or /bin/bash anyway, so as
long as bitbake can parse it you can use anything.

Ross

[-- Attachment #2: Type: text/html, Size: 1870 bytes --]

Re: what shell programming constructs can we count on in .bbclass files?

[Patrick Ohly]

On Wed, 2017-03-15 at 11:05 +0000, Burton, Ross wrote:
> Our shell parser *should* be POSIX compliant.  There are a few places
> where it isn't, but I believe those should be considered bugs.  At the
> end of the day the shell is actually executed by /bin/sh or /bin/bash
> anyway, so as long as bitbake can parse it you can use anything.

Unfortunately bitbake has problems parsing useful things like $( ). Are
there open bugs about that and/or is it worth filing enhancement
requests?



-- 
Best Regards, Patrick Ohly

The content of this message is my personal opinion only and although
I am an employee of Intel, the statements I make here in no way
represent Intel's position on the issue, nor am I authorized to speak
on behalf of Intel on this matter.

Re: what shell programming constructs can we count on in .bbclass files?

[Robert P. J. Day]

On Wed, 15 Mar 2017, Patrick Ohly wrote:

> On Wed, 2017-03-15 at 11:05 +0000, Burton, Ross wrote:
> > Our shell parser *should* be POSIX compliant.  There are a few
> > places where it isn't, but I believe those should be considered
> > bugs.  At the end of the day the shell is actually executed by
> > /bin/sh or /bin/bash anyway, so as long as bitbake can parse it
> > you can use anything.
>
> Unfortunately bitbake has problems parsing useful things like $( ).
> Are there open bugs about that and/or is it worth filing enhancement
> requests?

  on a more basic level, there's quite a lot of, i guess, bourne
shell'isms (and other weirdnesses) still lying around. in
meta/classes/, some examples (hoping i didn't screw them up):

  "if test" rather then "if ["

  all that 'if [ "x${VAR} = "x" ]' i mentioned earlier

  if [ "x${ICE_PATH}" = "x" ]   rather than   if [ -z "${ICE_PATH}" ]

  if [ ! -z "${conf_sign_keyname}" ]  ... ?????? :-)

none of this is a big deal, but there sure are some historical
holdovers and strangely-chosen constructs.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

Re: what shell programming constructs can we count on in .bbclass files?

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 319 bytes --]

On 15 March 2017 at 12:40, Patrick Ohly <patrick.ohly@intel.com> wrote:

> Unfortunately bitbake has problems parsing useful things like $( ). Are
> there open bugs about that and/or is it worth filing enhancement
> requests?
>

Absolutely worth filing bugs.  Worst case, they get marked as duplicates.

Ross

[-- Attachment #2: Type: text/html, Size: 729 bytes --]