on_read & on-write config for ConcatenateBytes #186
Conversation
markmandel
added
kind/feature
There was a problem hiding this comment.
Concatenatebytes doesn't sound like it needs a direction 😯? Thinking prepending isn't the opposite of Appending so that If I append bytes in one direction, nothing suggests I would want to prepend or even mutate packets in the opposite direction? Also There could be more operations to insert bytes into a packet e.g we could have an insert bytes at offset x in each packet so that the current upstream/downstream would not be relevant and we won't be able to support that
|
But what if I want to add bytes going from server->client, not just client->server ? |
|
Just re-reading this and maybe the confusion is that it sounds like it allows the Filter to add bytes BOTH on data going upstream or downstream -- which is not the case. This only give you the option to append data going EITHER upstream OR downstream. Does that make more sense? |
|
This is likely also blocked by the conversation in #188 as well. |
|
From: #188 (comment) This should be changed to: |
markmandel
force-pushed
the
feature/direction-concat
branch
from
f3a6fd0 to
f96c52d
Compare
markmandel
changed the title
direction configuration for ConcatenateBytes
markmandel
force-pushed
the
feature/direction-concat
branch
2 times, most recently
from
d741d84 to
f34276f
Compare
|
This should be good for re-review, as the config structure and documentation is now inline with #188 (comment) |
| type: string | ||
| description: | | ||
| Either append or prepend the `bytes` data to each packet filtered on write of the listening port. | ||
| default: "DO_NOTHING" |
There was a problem hiding this comment.
| default: "DO_NOTHING" | |
| default: DO_NOTHING |
| default: "APPEND" | ||
| enum: ['APPEND', 'PREPEND'] | ||
| Either append or prepend the `bytes` data to each packet filtered on read of the listening port. | ||
| default: "DO_NOTHING" |
There was a problem hiding this comment.
| default: "DO_NOTHING" | |
| default: DO_NOTHING |
markmandel
force-pushed
the
feature/direction-concat
branch
from
f34276f to
b6b4da1
Compare
This allows for concatenating bytes both on_write and on_read - whereas previously you could only do on_read. This is also useful to implement an integration test that checks to see if correct Filter ordering happens both on listening port read and write.
markmandel
force-pushed
the
feature/direction-concat
branch
from
b6b4da1 to
e40b938
Compare
This allows for concatenating bytes both on_write and on_read - whereas previously you could only do on_read.
This is also useful to implement an integration test that checks to see if correct Filter ordering happens both on listening port read and write.