[BP] CLI/render: fix XFA document length recognition

Sample file:
https://www.canada.ca/content/dam/ircc/migration/ircc/english/pdf/kits/forms/imm5708e.pdf

src/pypdfium2/_cli/_parsers.py

@@ -82,8 +82,10 @@ def add_n_digits(parser):
     )
 
 
-def get_input(args, **kwargs):
+def get_input(args, init_forms=False, **kwargs):
     pdf = pdfium.PdfDocument(args.input, password=args.password, **kwargs)
+    if init_forms:
+        pdf.init_forms()
     if "pages" in args and not args.pages:
         args.pages = [i for i in range(len(pdf))]
     return pdf

src/pypdfium2/_cli/render.py

@@ -284,7 +284,7 @@ def main(args):
 
     # TODO turn into a python-usable API yielding output paths as they are written
 
-    pdf = get_input(args)
+    pdf = get_input(args, init_forms=args.draw_forms)
 
     # TODO move to parsers?
     pdf_len = len(pdf)
@@ -349,9 +349,6 @@ def main(args):
     if len(args.pages) <= args.linear:
 
         logger.info("Linear rendering ...")
-        if args.draw_forms:
-            pdf.init_forms()
-
         for i in args.pages:
             _render_job(i, pdf, kwargs, engine)
 

src/pypdfium2/_helpers/document.py

@@ -139,9 +139,8 @@ def init_forms(self, config=None):
         Initialize a form env, if the document has forms. If already initialized, nothing will be done.
         See the :attr:`formenv` attribute.
     
-        Note:
-            If form rendering is desired, this method should be called directly after constructing the document,
-            before getting any page handles (due to PDFium's API).
+        Attention:
+            If form rendering is desired, this method shall be called immediately after document construction, before getting document length or page handles.
         
         Parameters:
             config (FPDF_FORMFILLINFO | None):